Parking Pressure Identification using GIS Modelling

Smart cities are a result of constant technological advancements that aim to improve the lives of its residents. One of the most significant features of smart cities is urban mobility. Urban traffic congestion is becoming increasingly common in major cities as the number of vehicles increases. Additionally, drivers all around the world struggle to find parking spaces in urban areas. Parking search traffic causes increased travel times and air pollution. So, it is essential to provide real time parking availability to the drivers to reduce the traffic congestion and protect from the increased emissions that affect the environment negatively. So that in this thesis we have used the Internet-of-Things (IoT) technologies and other spatial data which helps to build a model for the parking occupancy minutes prediction in the city of Berlin using Random Forest (RF) Regression machine learning algorithm. The RF Regression model achieved R2 value of 83 % which represents the accuracy of the model. Using the build model from the IoT sensors the parking pressure regions for the areas without sensor are predicted. By identifying the regions with high parking pressure, we can install the IoT sensors which will provide real time parking availability to the drivers through the mobile application.